bci2000.org BBS

Dear Colleagues.

Hello. I am a new user of BCI2000. I use a cheap USB amplifier. Unfortunately, this was not supported by BCI2000. So, I just made a data acquisition module according to your kind instructions and it works well with BCI2000 core modules. I tested it with BCI2000 Launcher. Now, I am doing the P300 speller experiment. But I did not made success yet.

I have questions as belows;

1) Is the Source Channel Gain fixed to certain amplifier or variable with circumstances? When I use this parameter with 0.010 in my amplifier, the graph is very similar with examples but I do not exactly know this value is the exact mcV.

2) In P300 response, should the positive reflex be in a range of 250 to 550ms? In some cases, this value was in a range below 200ms.

3) How much sensitive P300 response is to experimental circumstances (i.e. eye blinking, muscle movement) ?

Thanks.

Sincerely,

Hello Cistern,

cistern wrote:Dear Colleagues.

1) Is the Source Channel Gain fixed to certain amplifier or variable with circumstances? When I use this parameter with 0.010 in my amplifier, the graph is very similar with examples but I do not exactly know this value is the exact mcV.

The source channel gain is dependent on your amplifier, and converts the values transmitted from the amplifier from A/D units to micro-V. This value depends on the range of the amplifier (+/- muV) and the resolution (in bits). The amplifier should have documentation that explains how to convert from A/D to muV.

2) In P300 response, should the positive reflex be in a range of 250 to 550ms? In some cases, this value was in a range below 200ms.

This can depend on which electrodes you are seeing the response on. If the response at 200 ms is on the occipital electrodes (O1, O2, Oz), it is likely a visual evoked response. Fortunately, this signal can also be used for the P300 task. I would recommend using the P300GUI or the P300 Classifier tools to generate a parameter file, which will automatically use the best features, including those features at 200 ms or less.

3) How much sensitive P300 response is to experimental circumstances (i.e. eye blinking, muscle movement) ?

In general, you always want to minimize these artifacts, obviously. However, if you are using many flashes (e.g., >10), these artifacts will be averaged out of the final ERP. However, this also implies that if you can minimize artifacts, then you will get a response using fewer flashes!

Let us know if you have other questions!
Adam Wilson